LOAM
LOAM 將 3D LiDAR 的同時定位與建圖拆成兩個並行、頻率不同的演算法：高頻（約 10 Hz）里程計（odometry）以掃描對掃描配準估計速度並校正運動畸變（motion distortion），低頻（約 1 Hz）建圖（mapping）再把去畸變點雲精細配準到地圖。兩者都只使用依局部平滑度挑出的邊緣點（edge）與平面點（planar），分別以點到線、點到面距離作為殘差，並以 Levenberg-Marquardt 最佳化求解。系統沒有迴圈閉合（loop closure），IMU 只是選用的前處理先驗，因此長距離漂移無法做全域修正。
本頁內容
LOAM splits lidar SLAM into a fast scan-to-scan odometry that also removes motion distortion and a slower scan-to-map refinement, both using only smoothness-selected edge and planar points with point-to-line/plane residuals; it has no loop closure.
技術屬性
欄位內容為文獻擷取紀錄的原文用語（英文），以原文為據；「未查證」表示本研究尚未讀到該資訊，不代表該方法不具備此能力。
| 感測輸入 | 3D LiDAR (custom rotating Hokuyo UTM-30LX 2D scanner)、IMU (optional, Xsens MTi-10)、3D LiDAR (360 deg Velodyne lidar via KITTI, 10 Hz; the model is not named in this paper) |
|---|---|
| 原文測試平台 | cart (pushed, indoor)、ground vehicle、handheld、vehicle (KITTI) |
| 狀態估計 | Levenberg-Marquardt nonlinear least squares with bisquare robust weights; scan-to-scan odometry at about 10 Hz and scan-to-map mapping at about 1 Hz running in parallel (Sec. IV-B, V-C, VI) |
| 資料關聯 | edge and planar feature points selected by local smoothness per scan line; point-to-line and point-to-plane distances; mapping stage finds line/plane correspondences by eigen-analysis of local map point clusters (Sec. V-A, V-B, VI) |
| 時間表示 | discrete sweep poses with constant angular and linear velocity (linear pose interpolation) inside a sweep (Sec. V-C) |
| 去畸變 | points reprojected with the linearly interpolated odometry pose; optional IMU preprocessing removes orientation change and part of acceleration-induced distortion (Sec. V-C, VII-B) |
| 迴圈閉合 | none (stated in Sec. I and listed as future work in Sec. VIII) |
| 全域最佳化 | none |
| 地圖表示 | registered point cloud map Q_k built from the undistorted sweeps, stored in 10 m cubes; cubes intersecting the new sweep are loaded into a KD-tree; matching uses ten times more feature points than odometry with edge or plane neighbourhoods found by eigen-analysis; the map is downsized with a 5 cm voxel grid |
| 先驗資訊 | none (IMU optional) |
| 可輸出幾何 | motion-corrected registered point cloud map (5 cm voxel-grid downsampled) and 6-DoF pose at about 10 Hz (Sec. VI); dense raw-point export not described in paper |
| 計算需求 | real time on a laptop with 2.5 GHz quad cores and 6 GiB memory; odometry and mapping each run on a separate core; ROS on Linux |
使用設備
原文使用的感測器、運算硬體與載具（equipment）。型號保留原文寫法，連結到設備頁中同一型號的歸併名稱；角色依原文用途分為方法輸入、資料集感測器、執行運算平台、參考或真值量測（reference or ground truth）與比較對象設備。
| 類別 | 型號（原文寫法） | 角色 | 資料集 | 原文規格 | 出處 |
|---|---|---|---|---|---|
| LiDAR | Hokuyo UTM-30LX (custom rotating 3D lidar)歸入：Hokuyo UTM-30LX | 方法輸入 | 未標示 | 180 deg FoV, 0.25 deg resolution, 40 lines/s; motor rotates at 180 deg/s between -90 and 90 deg (1 s sweep); encoder resolution 0.25 deg | (Zhang & Singh, 2014, Sec. IV-A; Fig. 2) |
| LiDAR | Velodyne lidar (360 deg, model not named) | 資料集感測器 | KITTI odometry benchmark | logged at 10 Hz | (Zhang & Singh, 2014, Sec. VII-C; Fig. 13) |
| 慣性量測單元（IMU） | Xsens MTi-10 | 方法輸入 | 未標示 | optional; orientation from a Kalman filter on gyro and accelerometer used to preprocess the point cloud | (Zhang & Singh, 2014, Sec. VII-B) |
| GNSS 接收器 | high accuracy GPS/INS | 參考或真值量測 | 未標示 | on the ground vehicle in the orchard drift test | (Zhang & Singh, 2014, Sec. VII-A) |
| GNSS 接收器 | high accuracy GPS/INS | 參考或真值量測 | KITTI odometry benchmark | ground truth of the KITTI benchmark | (Zhang & Singh, 2014, Sec. VII-C) |
| 載具平台 | cart pushed by a person | 方法輸入 | 未標示 | carries lidar, battery and laptop; indoor tests at 0.5 m/s | (Zhang & Singh, 2014, Sec. VII-A) |
| 載具平台 | ground vehicle | 方法輸入 | 未標示 | lidar mounted at the front; outdoor tests at 0.5 m/s | (Zhang & Singh, 2014, Sec. VII-A) |
| 載具平台 | handheld | 方法輸入 | 未標示 | person walks at 0.5 m/s moving the lidar up and down about 0.5 m; staircase test | (Zhang & Singh, 2014, Sec. VII-B) |
| 運算硬體 | laptop, 2.5 GHz quad cores | 執行運算平台 | 未標示 | 6 GiB memory; two cores used | (Zhang & Singh, 2014, Sec. VII) |
| 其他 | tape ruler | 參考或真值量測 | 未標示 | manual ground-truth measurement for the IMU comparison tests | (Zhang & Singh, 2014, Sec. VII-B) |
作者報告的優勢與限制
優勢
- Real-time odometry and mapping without high-accuracy ranging or inertial measurements (abstract; Sec. VII)
- KITTI odometry benchmark: authors report 0.88% average position error over 100-800 m segments and first rank at submission time (Sec. VII-C)
- Indoor relative drift about 1% and outdoor about 2.5% in the authors' corridor/orchard tests at 0.5 m/s (Sec. VII-A, Table I)
限制
- No loop closure, so accumulated drift is not corrected (Sec. I, VIII)
- Assumes smooth, continuous velocity within a sweep; an IMU is needed for abrupt motion (Sec. III, VII-B)
- Matching errors larger in natural outdoor scenes than in man-made indoor scenes (Sec. VII-A, Fig. 11)
- KITTI results are not reported in the paper itself, only the benchmark rank and 0.88% average error (Sec. VII-C)
- Own tests are short (27 to 67 m) at 0.5 m/s (Sec. VII-A, VII-B, Tables I and II)
- Follow-up work states LOAM's global voxel map makes loop closure and absolute measurements (e.g., GPS) hard to add and that it drifts in large-scale tests (Shan et al., 2020, Sec. I)
- Follow-up work reports that LOAM's feature extraction cannot keep up on an embedded Jetson TX2 and that noisy ground/vegetation features cause divergence on a small UGV (Shan & Englot, 2018, Sec. I, IV)
營建工程相關證據
原論文的室內測試在走廊與大廳（既有建築）中進行，並以固定站點掃描比對地圖匹配誤差，但未在營建工地驗證。Feng 等人（Feng et al., 2025）在西安醫院門診大樓施工現場（主體結構已封頂、轉入機電安裝與室內裝修階段）與依施工圖建立的 Gazebo 模擬工地中測試「LOAM」，並稱所有演算法使用官方開源儲存庫的預設參數；但 LOAM 原始官方程式已無法取得，該文 Sec. 4.1.1 僅提及 A-LOAM 為開源實作，未明言實際執行的版本（推論：可能為 A-LOAM）。作者報告 LiDAR-only 方法在 z 軸累積漂移並造成地圖翹曲（Sec. 5.2-5.3）。論文未說明實際工地 APE 所用參考軌跡的來源（全文僅描述 Gazebo 模擬的真實軌跡外掛），故實際工地 APE 只能視為作者報告值，不能當作已驗證的幾何精度。
原文驗證環境：公開基準、受控實驗、已完工建築、獨立參考量測
報告的性能數據
以下是原文作者報告的性能數值（author-reported results），不是本研究重新量測的結果。每張圖只並列同一個比較組（comparison group，同一張表、同一組實驗設定）內的方法；不同比較組之間的數值不可直接比較，也不構成排名。
本方法共出現在 49 個比較組，合計 311 筆紀錄。以下列出本方法紀錄最多的 4 組，其餘 45 組列在最後，並連到性能比較頁。
Shan & Englot, 2018 · Table IV 本方法 24 筆
表格設定（擷取紀錄原文）：Runtime of each module for processing one scan, averaged over 10 real-time trials; LOAM has no segmentation module (Shan & Englot, 2018, Table IV)
runtime of segmentation module per scan，Own Jackal UGV datasets · Experiment 1
只並列這張表在相同設定下報告的方法；以「本方法：」開頭者為本頁方法。失敗、未執行與未報告以標記呈現，不是 0。
- 不適用
按 Tab 進入圖表後，用上下方向鍵逐一瀏覽各類別，Esc 關閉提示框；也可開啟表格檢視閱讀全部數值。
這些是 Shan & Englot, 2018 在此表設定下報告的數值（author-reported results），只能在同一個比較組內對照，不代表方法在其他資料或設定下的表現。
資料來源作者報告值(Shan & Englot, 2018, Table IV)
| 方法（原文寫法） | 報告值 | 出處 |
|---|---|---|
| LOAM（Nvidia Jetson TX2 (ARM Cortex-A57), CPU only）本方法 | 無數值不適用註記（擷取紀錄）：不適用 (N/A in table) | (Shan & Englot, 2018, Table IV) |
| LOAM（laptop, 2.5 GHz Intel i7-4710MQ, CPU only）本方法 | 無數值不適用註記（擷取紀錄）：不適用 (N/A in table) | (Shan & Englot, 2018, Table IV) |
| LeGO-LOAM（Nvidia Jetson TX2 (ARM Cortex-A57), CPU only）原文提出 | 29.3 ms | (Shan & Englot, 2018, Table IV) |
| LeGO-LOAM（laptop, 2.5 GHz Intel i7-4710MQ, CPU only）原文提出 | 16.7 ms | (Shan & Englot, 2018, Table IV) |
Pan et al., 2021 · Table II 本方法 15 筆
表格設定（擷取紀錄原文）：KITTI odometry ATE [%] and ARE [deg/100m], averaged over 100-800 m segments; baseline values from original papers and KITTI leaderboard; * = with loop closure; time in s per frame (Pan et al., 2021, Table II)
ATE [%] (average translation error)，KITTI odometry · 00-10 mean
只並列這張表在相同設定下報告的方法；以「本方法：」開頭者為本頁方法。失敗、未執行與未報告以標記呈現，不是 0。
按 Tab 進入圖表後，用上下方向鍵逐一瀏覽各類別，Esc 關閉提示框；也可開啟表格檢視閱讀全部數值。
這些是 Pan et al., 2021 在此表設定下報告的數值（author-reported results），只能在同一個比較組內對照，不代表方法在其他資料或設定下的表現。
資料來源作者報告值(Pan et al., 2021, Table II)
| 方法（原文寫法） | 報告值 | 出處 |
|---|---|---|
| LOAM [10]本方法 | 0.84% | (Pan et al., 2021, Table II) |
| IMLS-SLAM [11] | 0.52% | (Pan et al., 2021, Table II) |
| MC2SLAM [13] | 0.52% | (Pan et al., 2021, Table II) |
| S4-SLAM [26]* | 0.92% | (Pan et al., 2021, Table II) |
| PSF-LO [27] | 0.74% | (Pan et al., 2021, Table II) |
| SUMA++ [16]* | 0.7% | (Pan et al., 2021, Table II) |
| LiTAMIN2 [51]* | 0.85% | (Pan et al., 2021, Table II) |
| LO-Net [18] | 0.83% | (Pan et al., 2021, Table II) |
| FALO [25] | 1% | (Pan et al., 2021, Table II) |
| LoDoNet [28] | 1.27% | (Pan et al., 2021, Table II) |
| MULLS-LO(mc)原文提出 | 0.49% | (Pan et al., 2021, Table II) |
| MULLS-SLAM(mc)*原文提出 | 0.52% | (Pan et al., 2021, Table II) |
| MULLS-LO(s1)原文提出 | 2.57% | (Pan et al., 2021, Table II) |
| MULLS-SLAM(m1)*原文提出 | 0.77% | (Pan et al., 2021, Table II) |
| MULLS-SLAM(m5)*原文提出 | 0.6% | (Pan et al., 2021, Table II) |
| MULLS-SLAM(s5m5)*原文提出 | 0.61% | (Pan et al., 2021, Table II) |
Zuo et al., 2020 · Table VI 本方法 14 筆
表格設定（擷取紀錄原文）：Averaged ATE of 5 runs on 6 Vicon-room sequences (cluttered room, Vicon ground truth), orientation (deg) and position (m); ATE computed following Zhang and Scaramuzza [23]; '-' = translational error above 20 m. The Average column is printed by the authors (for LIO-MAP and LIC-Fusion it averages only the sequences that did not fail). (Zuo et al., 2020, Table VI)
averaged ATE, orientation (deg)，Vicon Room sequences (authors' data) · Seq 1 (42.62 m)
只並列這張表在相同設定下報告的方法；以「本方法：」開頭者為本頁方法。失敗、未執行與未報告以標記呈現，不是 0。
- 失敗
按 Tab 進入圖表後，用上下方向鍵逐一瀏覽各類別，Esc 關閉提示框；也可開啟表格檢視閱讀全部數值。
這些是 Zuo et al., 2020 在此表設定下報告的數值（author-reported results），只能在同一個比較組內對照，不代表方法在其他資料或設定下的表現。
資料來源作者報告值(Zuo et al., 2020, Table VI)
| 方法（原文寫法） | 報告值 | 出處 |
|---|---|---|
| LIC-Fusion 2.0原文提出 | 2.537 deg | (Zuo et al., 2020, Table VI) |
| OpenVINS-IC | 2.625 deg | (Zuo et al., 2020, Table VI) |
| Proposed-LI | 2.333 deg | (Zuo et al., 2020, Table VI) |
| LOAM本方法 | 5.88 deg | (Zuo et al., 2020, Table VI) |
| LIO-MAP | 無數值失敗註記（擷取紀錄）：failed | (Zuo et al., 2020, Table VI) |
| LIC-Fusion | 2.345 deg | (Zuo et al., 2020, Table VI) |
Wang et al., 2021c · Table 1 本方法 13 筆
表格設定（擷取紀錄原文）：KITTI odometry, trained on 00-06 (marked *) and tested on 07-10; trel = average translational RMSE (%) over 100-800 m subsequences; rows other than LOAM w/o mapping and Ours are copied from LO-Net [10]; LOAM is a full system with mapping, others are odometry only (Wang et al., 2021c, Table 1)
trel (average translational RMSE, %)，KITTI odometry · 07 (test)
只並列這張表在相同設定下報告的方法；以「本方法：」開頭者為本頁方法。失敗、未執行與未報告以標記呈現，不是 0。
按 Tab 進入圖表後，用上下方向鍵逐一瀏覽各類別，Esc 關閉提示框；也可開啟表格檢視閱讀全部數值。
這些是 Wang et al., 2021c 在此表設定下報告的數值（author-reported results），只能在同一個比較組內對照，不代表方法在其他資料或設定下的表現。
資料來源作者報告值(Wang et al., 2021c, Table 1)
| 方法（原文寫法） | 報告值 | 出處 |
|---|---|---|
| Full LOAM [31] | 0.69% | (Wang et al., 2021c, Table 1) |
| ICP-po2po | 5.17% | (Wang et al., 2021c, Table 1) |
| ICP-po2pl | 1.55% | (Wang et al., 2021c, Table 1) |
| GICP [19] | 0.64% | (Wang et al., 2021c, Table 1) |
| CLS [21] | 1.04% | (Wang et al., 2021c, Table 1) |
| Velas et al. [22] | 1.77% | (Wang et al., 2021c, Table 1) |
| LO-Net [10] | 1.7% | (Wang et al., 2021c, Table 1) |
| DMLO [11] | 0.73% | (Wang et al., 2021c, Table 1) |
| LOAM w/o mapping (published code run by authors)本方法 | 10.87% | (Wang et al., 2021c, Table 1) |
| Ours (PWCLO-Net)原文提出 | 0.6% | (Wang et al., 2021c, Table 1) |
其他比較組
列出其餘 45 個比較組
- Ruan et al., 2023 · Table II
- Lv et al., 2021 · Table II
- Liu et al., 2023b · Table II
- Liu et al., 2023b · Table V
- Shan & Englot, 2018 · Table V
- Ye et al., 2019 · Table I
- Tagliabue et al., 2021 · Table 1
- Zhang & Singh, 2014 · Table II
- Yokozuka et al., 2021 · Table III
- Chen et al., 2022a · Table III
- Qin et al., 2020 · Table I
- Shao et al., 2019 · Table I
- Koide et al., 2021a · Table I
- Zhao et al., 2021 · Table II
- Zou et al., 2022 · Table IX
- Zuo et al., 2020 · Table V
- Feng et al., 2025 · Table 3
- Feng et al., 2025 · Table 4
- Wang et al., 2023b · Table IV
- Wang et al., 2023b · Table V
- Zuo et al., 2019 · Table II
- Zhang & Singh, 2014 · Table I
- Zou et al., 2022 · Table III
- Zou et al., 2022 · Table V
- Zou et al., 2022 · Table VI
- Frosi & Matteucci, 2022 · Table II
- Frosi & Matteucci, 2022 · Table III
- Frosi & Matteucci, 2022 · Table IV
- Lee et al., 2024b · Table 4
- Qian et al., 2021 · Table IV
- Zou et al., 2022 · Table IV
- Liu & Zhang, 2021 · Table I
- Liu & Zhang, 2021 · Table II
- Zuo et al., 2019 · Table I
- Qian et al., 2021 · Table V
- Yan et al., 2026a · Table 2
- Zou et al., 2022 · Table II
- Liu & Zhang, 2021 · Text Sec.VI-A
- Deschaud, 2018 · Text Sec. VI-B
- Yokozuka et al., 2021 · Table III (overall)
- Yokozuka et al., 2021 · Table IV
- Yokozuka et al., 2021 · Table V
- Zhang & Singh, 2014 · Text Sec. VII-C
- Wang et al., 2021b · Text Sec. IV-B
- Zhang & Singh, 2018 · Table 4
來源
Zhang & Singh, 2014
（2014）LOAM: Lidar Odometry and Mapping in Real-timeRobotics: Science and Systems X (RSS 2014), RSS X, paper p07 (proceedings URL rss10/p07.pdf)
同儕審查已出版已讀全文經典查證後修正
相關版本
- 期刊延伸版：Low-drift and real-time lidar odometry and mapping 10.1007/s10514-016-9548-2
- 程式碼釋出：loam_back_and_forth / loam_continuous (ROS wiki pages cited in paper; linked GitHub repositories return 404 on 2026-09-25) http://wiki.ros.org/loam_back_and_forth